Paper Title
Computed Tomography-Based Lung Cancer Detection Using Parallel Deep Convolution Neural Network

Abstract
Globally, lung cancer is the primary cause of death. The Lung Cancer Detection (LCD) is critical in the initial stages because of the asymptomatic nature and complex structure of the lung computed tomography (CT) images. Numerous deep learning (DL)-based plans have emerged in recent years, significantly improved LCD efficiency. However, the effectiveness of the DL-based schemes is limited because of complicated network topology, less generalization capability, poor feature representation, and class imbalance problems. This article presents LCD based on Parallel Convolution Neural Networks (LC-PCNN) to improve the feature distinctiveness and generalization capability. Further, it utilizes data augmentation (DA) based on rotation, scaling, shifting, and noise addition to minimize the problem of class imbalance. The outcomes of the proposed LC-PCNN are estimated on the Iraq-Oncology Teaching Hospital/National Center for Cancer Diseases (IQ-OTH/NCCD) Lung Cancer Dataset. The LC-PCNN offers an accuracy of 99.14%, recall of 0.99, precision of 0.99, F1-score of 0.99, selectivity of 0.99, also NPV of 0.99 for the augmented dataset, which is superior over traditional LCD methods. Keywords - Biomedical Image Processing, Computed Tomography, Deep Convolution Neural Network, Deep Learning, Data Augmentation